We want to work on enhancing the analytics for Risk Analysis.  Currently we comment when we see repeated failures that have high historical pass rates, however when a regression comes in from another PR we will flag that regression as a risk for each PR that sees the failure.

For this story we want to persist potential regressions detected by risk analysis in big query.  A potential place to insert logic is in pr_commenting_processor.

We want to make sure we can track the repo, pr, test name, potentially testid and associated risk.

Future work will include querying this data when building the risk summary to see if any tests flagged as risky within the current PR are also failing in other PRs indicating this PR is not the contributing factor and a regression has potentially entered the main branches / payloads.